Our review of the literature as well as our own work indicates that the probability of lethal disease occurring in otherwise normal animals when stressed is low; however, when an animal has some predisposition to disease and is then subjected to stress, lethal disease occurs much more consistently. Using this concept, we have predisposed guinea pigs to arrhythmias by using digitalis. When our digitalized animals are then stressed, their cardiac rhythm changes from a normal one to an abnormal potentially lethal one. This then is a good animal model of psychosomatic digitalis-toxicity. This model is an important one because digitalis is the fourth most commonly prescribed drug in the U.S. and has a low therapeutic:toxic ratio. Environmental factors have never been seriously considered as a reason for individual variability in response to digitalis. We are currently improving our model by using a free-ranging rather than a restrained subject and by using IV administration of our cardiac glycoside, ouabain, rather than IP. Our objectives are to use this improved model to learn the relative importance of the autonomic nervous system in the production of the arrhythmias; we will do this by pharmacological and surgical ablations and by neurophysiological recording and stimulation; we also plan to carry these experiments into the brain in an attempt to locate sites responsible for the psychosomatic effect. Another objective is to begin to analyze whether specific emotional states are responsible for the effect or merely a generalized arousal. Finally, we shall explore how manipulation of psychological contingencies can protect the animal from the lethal outcome; this will be done (a) by manipulating controllability over shock and by training the subject to maintain a constant heart rate by operant techniques.